Back to 2014 Annual Meeting Posters

---

Improvement in Pancreatectomy Outcomes: Lessons Learned From Applying Evidence-Based Practice
Iryna Chesnokova^*, Benjamin Johnson, Anna M. Button, Kelly Petrulevich, Howe R. James, James J. Mezhir
Surgery, Univ of Iowa, Iowa City, IA

Background: Morbidity following pancreatectomy remains significant despite improvements in patient selection and operative mortality. We set out to determine the impact of an evidence-based and specialized multidisciplinary program including surgeons, residents, anesthesiologists, and nurses on outcomes following pancreatectomy.
Methods: A retrospective analysis of prospective institutional pancreatectomy and ACS-NSQIP databases were reviewed to identify patients who had undergone elective pancreatectomy from 9/2007 to 12/2012. The initiation of our multidisciplinary program in 2010 included judicious perioperative fluid management and transfusion practices, evidence based drain and catheter management, and an orientation and teaching program for nursing staff to facilitate standardized postoperative patient protocols. Multivariate linear or logistic regression was used where appropriate to determine if there was an improvement in operative outcomes over time after adjusting for significant patient-specific variables. Endpoints of interest included operative time, length of stay, rate of ICU admission, transfusion rate, and incidence of operative complications.
Results: Complete information was available for 226 patients who underwent pancreatectomy during the time period analyzed, including pancreaticoduodenectomy/total pancreatectomy (n=147) or distal pancreatectomy (n=79). There were 160 patients who experienced a complication (morbidity rate=70.8%) and 8 patients died within 90 days (mortality rate=3.5%). To control for patient-specific variables, multivariate analysis was performed for each individual operative outcome. When controlling for these factors, operative year was an independent predictor of improved outcomes - reduced operative time, length of stay, ICU admissions, transfusions, and complications. When adjusting for significant preoperative and intraoperative variables over time, there remained a significant improvement in all endpoints (Table).
Conclusions: Following the implementation of an evidence-based pancreatectomy program, there was a significant reduction in operative time, length of stay, the rate of ICU admission, blood transfusion, and complications. Multidisciplinary programs using evidence-based guidelines can help reduce operative complications and improve outcomes in pancreatectomy.

Table. Logistic regression analysis of changes in operative outcome measures over time.

Year	Number of Cases	Mean Operative Time (min)*	Mean Length of Stay (days)*	ICU Admission (%)	Complication Rate (%)	Transfusion Rate (%)
2007	14	490.5	24.7	50	85.7	71.4
2008	37	455.7	23.0	27	78.4	48.7
2009	47	420.9	21.3	32	76.6	40.4
2010	38	386.1	19.6	26	65.8	34.2
2011	44	351.3	17.9	20	65.9	45.5
2012	46	316.5	16.2	17	63.0	10.9
p-value		<0.0001	<0.0001	<0.0001	0.022	0.033

*Means are adjusted from multivariate linear regression model to be independent of patient-specific factors **p-values from multivariate models for year of operation

---

Back to 2014 Annual Meeting Posters